Dylan Oliva

I'm Dylan. Here's how I generated millions for home service businesses, saved them thousands of hours, and worked with brands doing over $100 million a year. Where it all started - 2022 - One day I was sitting in the library, finished my homework, and was watching YouTube on how to make money with marketing. Everyone needs it, would be a good skill to learn. Right after taking notes and writing it all down on a single sheet of paper it was time to execute. I built a website. Started asking around telling people about it. No clue what I was doing. Talked to a friend, he wanted to partner, we did. Got a roofing client. Did not know what we were doing and he eventually stopped paying us. That was when I realized I never want to underdeliver again. Always want to show value or you don't pay. I don't want to scam people. I want to make money honestly. So I dove deep into learning it all. Websites, SEO, ads, and google business profile. 2023 - Now that I was confident in my skill I was cold calling, trying to get clients. Got some through word of mouth. Was working with a music artist @YungLordFiness, a hair salon, a real estate guy, and a media guy who was referring me roofers. I started niching down into home services. Also probably in the best shape of my life here. Going to jujitsu twice a week and really counting my calories. 2024 - Helped generate over $5 Million for local businesses. Got an offer at @UBS. Accepted it. Thought I might learn something. Started a Pickleball club I started posting content on YouTube. Just sharing the things I learned over the years, and how I was getting results. 2025 - I started posting marketing content on Twitter, YouTube, and LinkedIn. Then I started the job @UBS. The agency was on the back burner but I was growing on Twitter and the community was great so I kept posting. Started a remote cleaning company. 30 days in, consistent leads, ranking top 3 on google. Quit @UBS 3 months in absolutely sucked. Joined a startup in the home services space on the founding team. Sold my marketing agency to a friend and went all in. We were working 8am to 11pm. Sometimes even later. Waking up even earlier. Traveling every week. Learning so much. Truly a great time, great people, smart people. I was working on home services companies some doing over $100 million a year. Implementing software for them. Helping build business cases to present to private equity groups. Constantly putting out fires, training AI agents and setting them up, managing and consulting customers. Working with the highest level home service companies. Trained for a half marathon. My plan initially was work there for 6 months and then do sales for 6 months. 2026 - I started posting on Twitter again. I was 4 months in, starting to build myself up again. Thinking about the transition to sales. Month 6 I left. Got an offer to work in sales at another startup. But I was thinking, I can do this. I know exactly what to do now. So I went all in! Today: Started Bloomview Growth. Client: added 628K in new revenue. No new hires. Response time went from 4 hours to under 5 minutes. 40 Google reviews to 180. Lead capture went from 34% to 58% on the same ad spend. Another client built an automated follow-up sequence. Closed 141K in 90 days. Built a Roofing AI Agent to find discrepancies in insurance supplements (filings), as well as a permit tracking agent to always stay up to date and file permits for jobs. Now I work on businesses end to end, from when a lead comes in all the way to when you're doing the job and forever after. If you made it this far going to put my 5 step framework on how I audit businesses in this thread.

@_toddanderson Todd could we jump on a call ?

Quoted three dev jobs last quarter at the same numbers I would've quoted two years ago for near six figures. The assumption is that LLMs ruined software pricing. They didn’t…. Really, they ruined typing code, which was already not the hard part. The hard part is creativity, and architecture familiarity. Security models. Hosting. Database choice. Persistence layer. Unit tests. Knowing which decisions you can walk back later and which ones you can't. A vibe-coded app from a stranger on Twitter goes to production with hardcoded keys and a SQL injection vector. A real dev shop hasn't gotten cheaper because the work that prevents that hasn't gotten easier. Most owners I quote understand this within ten minutes of trying to build their own app right now by themselves.

Helped a roofer rank for "metal roofing" in his city in 60 days. Average position went from 13 to 6. But we 4x his business. Pretty standard SEO results honestly. Any decent agency can move a needle on rankings. Here's what nobody tells you. The ranking wasn't the win. The win was what happened after the phone started ringing. Before we touched anything his team was responding to leads in 3 to 4 hours. Estimates went out and sat in a folder. No follow up sequence. No reactivation. Job gets done and nobody asks for a review. Past customers never hear from him again. So we ranked him. Calls went up. But revenue barely moved. Because the leads were leaking out the back just as fast as they were coming in the front. Here's what actually 4x'd the results from the same rankings: Built a follow up sequence on every estimate. Day 1, 7, 14. Different message each time. Not "just checking in." Actual value. Updated pricing. Job photos from similar work. Urgency when it made sense. Lead capture tightened up. Every inquiry gets a response in under 2 minutes. Not 2 hours. 2 minutes. Qualification happens automatically so the sales guy only talks to real buyers. After the job: automatic review request with a direct link. Not "please leave us a review." A link that opens the box. Past customers get re-engaged every 60 days. Seasonal offers. Maintenance reminders. Referral asks. Outbound on top. Built a list of homeowners in the service area who matched the ideal job profile. AI runs a 4 touch sequence over 14 days. First touch is value. Not spam. The rankings brought in the leads. The systems turned them into revenue. Most agencies stop at the rankings and send you a report. That's maybe 20% of the job. The other 80% is what happens after someone finds you. Truth is: if your agency got you to page 1 and revenue didn't change much, the problem isn't the rankings. It's everything that happens after the click.

The dumpster fire isn't the work itself. It's doing work a system should handle while the real money-making work piles up. Google brings the leads. AI catches them, follows up, gets the review, and re-engages past customers. The team goes back to only doing the jobs. That's when the sticker comes down.

Flew out to a client's shop last week. First thing I saw on the office manager's desk was the classic dumpster fire. She laughed and said "yeah that's pretty much every day around here." Calls coming in that nobody catches. Estimates sitting in folders. Reviews never asked for. The same 12 questions answered 40 times a day. Calendar chaos. Techs walking in blind because the notes are wrong. Every home service office I walk into has some version of this eraser. Or an sticker. Or a mug. Everyone knows it's a mess. By day 3 we had AI handling 70 to 80% of the admin that created that feeling. Follow ups automated. Reviews going out after every job. Leads getting texted back in 90 seconds. Calendar updates running on their own. The boring work handled. She told me the eraser might need to come down soon. That's the whole job. Turn the dumpster fire into a system.

A plumbing shop I work with had over 400 estimates sitting completely cold in their CRM. Most of them written off as lost. On top of that they had zero outbound. No proactive outreach. No follow up cadence. Just waiting for the phone to ring. Truth is: the money was sitting in two places at once. The leads they already paid to get. And the homeowners in their market who matched their ideal job but never called. They were only going after one of them. Here's what nobody tells you. Most home service businesses doing 1M to 3M run on inbound only. Lead comes in. You quote. You hope they call back. That's not a system. That's a prayer. We built one AI pipeline that handled both sides. Inbound recovery first. Every new estimate got an AI text within 90 minutes. Then 3 follow ups over 10 days that sounded human not robotic. Any reply with "maybe" or "thinking about it" got flagged for the owner to call personally. Every estimate older than 14 days got pulled into a reactivation sequence automatically. Then the outbound side. The part most people skip. Built a clean list of every homeowner in the service area who matched their ideal job profile. Home age. Square footage. Past service history from public data. AI ran a 4 touch sequence over 14 days. First message referenced a common issue in their neighborhood. Second was a seasonal offer tied to current weather. Third was social proof with job photos from similar homes nearby. Fourth was a low pressure close. None of it sounded like spam. Every message added something real. 30 days later. 141k closed from dead estimates alone. Outbound added another 47 booked calls from homeowners who never would have called on their own. What I learned: most home service businesses doing 1M to 3M have the same two leaks. They're not recovering the leads they already paid for. And they're not going after the ones sitting in their backyard. When both run together the business stops feeling like it's at the mercy of whoever calls today.

A home service business I work with was stuck at 1.2M for three straight years. Same crew size. Same trucks. Same market. No big marketing budget increases. After we installed the full AI pipeline they hit 1.8M in the next 12 months. Here's exactly what changed. The owner wasn't lacking leads. He was losing too many of the ones he already had and wasting too much of his team's time on admin. We mapped the entire operation and built one connected system that attacked both problems at once. Speed to lead and qualification. Every inbound lead now gets an AI text within 90 seconds. Qualifies the job and books the appointment on the spot when possible. No more "I'll get back to you" texts that die in the customer's inbox. Dead estimate recovery. Anything that goes cold gets pulled into an automatic sequence. Updated pricing. Job photos. Urgency triggers. No manual chasing required. Team time reclamation. We documented every repetitive office task. Review requests. Follow up texts. Calendar updates. Basic customer questions. Built AI agents that handle 70 to 80% of it exactly the way the best CSR already does. The team only touches the high value work now. The numbers after 12 months: Average jobs per truck per week went from 4.8 to 6.2. Close rate on estimates jumped 15 percentage points. The office team went from drowning to having breathing room. Revenue increased 600k with zero additional hires. Same techs on the road. Same trucks. Same owner. The business just stopped leaking money and time in the middle of the pipeline. Truth is: most 1M to 3M local service owners think growth means "get more calls." The faster path is usually making the calls you already get turn into jobs and giving your team their time back so they can actually run those jobs. When those two things run together the revenue follows without burning everyone out.

@Bettywayy Sending that over

@DylanOlivaGrows AI

This is exactly what I've been doing, except I'm not charging $2-3k/month. I built the full AI stack for my own cleaning company. Freed my time up from 20hrs a week to 5hrs. Margins went from 20% to 40%. Now I am doing this for other businesses for free to show I can bring value first before even asking for their credit card. Giving away the full playbook for free so you can see what's actually inside: Like this post & comment "AI" and I'll DM it to you. Here's what you're getting: • How to audit which roles in your business AI can actually handle • How to set up a custom agent trained on your processes (not a generic GPT wrapper) • How to connect it to your CRM, Slack, and reporting tools • Sub-agents for onboarding, client comms, and data tracking The "AI guy" model works. But most people teaching it have never actually run a business with these systems. I have and built agents out for home service businesses doing over $100 million a year. Drop "Agent" below.

Met a landscaper who pulls 40+ reviews a month doing one weird thing. 7 days after every job, he has a college student drive back to the customer's house and hand-deliver a handwritten thank-you card. Says it was a pleasure working with them and asks for a review on their experience. The same college student also handles his door hanger campaigns in the surrounding neighborhoods on the same route. His response rate on the cards is over 40%. Most automated SMS campaigns top out at 10-15%. The college kid costs him less than $20/hour. His average ticket is $2.4K. Getting reviews consistently takes hard work, but it pays off when you hold your position in the Map Pack.

The opportunity Grant's describing is real. I do this right now for home service companies. I have 2x companies using this and saved them hundreds of hours of time. But heres what he's not telling you. You don't just bring three AI platforms into a company and collect a check. The AI doesn't respond correctly at first. Its not trained right. It says the wrong things. Every industry has edge cases and weird scenarios that no platform handles out of the box. A plumbing company and a roofing company sound similar until you actually sit inside both of them. The workflows are completely different. The customer conversations are completely different. And when the AI messes up it doesn't just look bad. It actually impacts the business. You have to call the client and explain why their AI just quoted a customer the wrong price on a 12,000 dollar job. You have to fix it. You have to make sure it doesn't happen again. And if it does you have to own that too. Thats the real work. Not the software. The tweaking. The training. Sitting in the building watching how the team actually works and adjusting the system until it fits. I personally fly to every client and onboard them in person because you cannot do this over a zoom call. The opportunity is real but its messier than he makes it sound. Its operational. And thats exactly why most people wont last in this space. If you actually want to see how this works I put together a full breakdown on how I build this for companies. What I included: 1.) What roles I replaced with AI and what I kept human. 2.) The exact agents I built and what each one handles. 3.) Where I messed up and what I had to rebuild. 4.) The cost breakdown of what I was paying people vs what I pay now. And how I onboard a new client from first call to fully live system. Comment "SAAS" and I'll send it to you.

@scottmartinis Sending over now

@DylanOlivaGrows Saas

@Elliott30128705 Sending over now

@DylanOlivaGrows Saas

This is exactly what I do. I run a cleaning company and built AI agents to run most of the operations for me. Answering calls. Following up on quotes. Rebooking past customers. Sending review requests. All automated. Went from 20 hours a week running it to 5. Margins went from 25% to 60%. The business itself is simple. Cleaning, power washing, plumbing, HVAC. None of it is complicated. The edge isnt the service. The edge is using AI to run it while everyone else is still doing everything manually. Most home service companies are still paying someone 40-50k a year to answer phones and follow up on leads. AI does that for 200-500 a month and doesnt miss calls at 9pm on a Friday. The boring business is the vehicle. AI is the engine. Pick any of these 8. Build it. Then automate the parts that dont need a human. Thats how you print money in 2026.

@therobertbrooks This is incredible growth especially the last few months and I’ve looked at like 100 different service titan dashboards. Do you attribute this to the flat OPEX or what changed ?

April is officially over: 509K (!!!) in completed revenue - 174% growth YOY Gross Margin notched up a few points to over 50% EBITDA will land above 30% for the month - massive revenue growth with flat OPEX helps a ton. On to the next one

@RealPlantBrah Life of a arborist

I love it when my drop zone is a roof 😂

@irentdumpsters @remotecleanguy Bodhi your hairline is super dialed hella respect to that

chilling with @remotecleanguy my dawg gonna clear 2 mil this year in house cleaning 🔥

I was at a plumbing shop this week onboarding a customer. Was talking to the owner and the sales tech. The sales tech had been going back and forth with a lead and asks the owner what should I offer them. The owner sighed and just said call the guy right now and say man what is the number its going to take to close this contract. Thats it. No fancy proposal. No 47 slide deck. No "let me circle back." The sales tech overthought it for 3 days and the owner solved it in one sentence. I believe the straightforward honest true value approach is always the way. I genuinely want to provide value to my clients and build a partnership not just close a deal. Thats why I travel to each one and onboard them in person. You cant build a real partnership over a zoom call and a PDF. You build it by showing up and saying what is the real problem and how do we fix it together.

Live local SEO audit I did for @jamesonhaslam and West Coast Deck. He's ranking better than most deck builders I've audited, and almost all of it is coming from their killer reviews. Here are the changes I'd make to his profile that I think could 3x the decks they book from Google search:

@willametteshark SlopMaxing

@DylanOlivaGrows AI slop about AI slop.

@_toddanderson @Charles_SEO That awesome I love that

@DylanOlivaGrows Couldn't agree more. @Charles_SEO has always said SEO is an art not a science, I quote it in my sales meetings all the time. Art can't be mass produced, its gotta be bespoke and custom.

Had a sales call with a franchise recently for SEO. Same problem every franchise has… He can’t touch the website, corporate controls the page, and he has zero say in his local SEO. And, of course, the SEO is bad. Most agencies tell franchisees they’re stuck. What we built instead: Private rank and rent set up outside of the franchise system. He owns the GBPs, he owns he site, and corporate isn’t in the loop. He spun up a bunch of GBPs from the playbook in the first couple weeks. Started getting his own lead volume that he routes back to his franchise location. Franchisees aren’t stuck, they’re just told they are.